Optimal decentralized coordination control method for microgrid in new energy power system

Abstract

The present application relates to the technical field of large-scale power grid security, in particular to an optimal decentralized coordination control method for microgrid in new energy power system, comprising the following steps: based on the online obtained power grid topology and real-time line flow data, the virtual disconnection test is carried out on the key transmission line one by one, and the whole network cascade risk quantitative index is established. The present application realizes the simulation and analysis of line fault scene by real-time acquisition of power grid topology and line flow data, implementation of virtual disconnection test on key transmission line, establishment of whole network cascade risk quantitative index on this basis, and real-time comparison with safety threshold, identification of the initial fault line which is the biggest threat to system security, generation of vulnerable line risk contribution list, realization of positioning and advance prevention of high-risk line; further, the power adjustment amount required by microgrid group is directly calculated according to the risk contribution list, and the prospective flow diversion instruction set is formed.

Description

Technical Field

[0001] This invention relates to the field of large-scale power grid security technology, and in particular to an optimal decentralized coordinated control method for microgrids in a new energy power system. Background Technology

[0002] The field of large-scale power grid security technology mainly involves real-time monitoring, risk quantification assessment, and defense and control technologies for the operation status of large-scale, highly complex power grids under sudden accidents, harsh environments, or extreme operating conditions during power system operation.

[0003] While existing technologies possess the capability for real-time monitoring, risk assessment, and defense and control of sudden power grid accidents and extreme environments, they lack refined simulation and targeted management of critical line faults. This makes it difficult to effectively capture the true propagation path of cascading faults, resulting in a reactive approach to line faults rather than proactive risk control. Furthermore, existing technologies lack a multi-microgrid collaborative optimization and management mechanism considering renewable energy grid integration, and suffer from response lag and insufficient constraints in dynamic frequency support scheduling. This leads to frequent occurrences of energy storage exceeding limits, generator ramp-up rates exceeding standards, and line overloads in actual operation. Therefore, improvements are needed. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies and propose an optimal decentralized coordinated control method for microgrids in new energy power systems.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: an optimal decentralized coordinated control method for microgrids in a new energy power system, comprising the following steps:

[0006] Based on the power grid topology and real-time line power flow data acquired online, the key transmission lines are virtually disconnected one by one to establish a network-wide cascaded risk quantification index.

[0007] Based on the network-wide cascaded risk quantification index, a numerical comparison is made with the safety threshold. When the network-wide cascaded risk quantification index exceeds the threshold, the initial fault line with the highest contribution to the network-wide cascaded risk quantification index is identified, a vulnerable line risk contribution list is generated, and the power injection or absorption adjustment amount executed by the microgrid group is calculated and determined based on the vulnerable line risk contribution list, and the microgrid group's preventive power flow guidance instruction set is obtained.

[0008] During the execution of the microgrid group preventive power flow guidance instruction set, each microgrid unit continuously monitors and generates a cluster over-limit status information set. The cluster over-limit status information set and the instruction update amount issued by the cluster coordinator are compared with a preset threshold to generate a hierarchical event trigger activation signal.

[0009] Based on the hierarchical event-triggered activation signal, each cluster coordinator initiates collaborative optimization calculations to generate a cluster-level optimal control sequence. The cluster-level optimal control sequence is then checked and allocated to establish a microgrid dynamic frequency support scheduling scheme.

[0010] Preferably, the steps for obtaining the network-wide cascading risk quantification indicator are as follows:

[0011] Based on the power grid topology information collected online and the power flow values ​​of each transmission line in the current state, a simulation of disconnection operation is performed on any one line in the set of key transmission lines. The power flow change values ​​of the remaining lines are recorded, and the total power flow amplitude under the accident scenario is calculated by combining the power flow values ​​before disconnection. A table of disconnected line numbers and power flow comparison of each line after the accident is generated.

[0012] Based on the disconnected line number and the power flow comparison table of each line after the accident, the power flow increment and initial load power before and after the accident are extracted, and the load rate of each line in the current disconnection scenario is calculated. At the same time, the proportion of the initial power flow in the total power flow of the entire network is extracted, and a set of risk contribution values ​​of all lines in the disconnection scenario is generated.

[0013] Based on the risk contribution value set of all lines under the aforementioned disconnection scenario, calculate the network-wide cascading risk quantification index.

[0014] Preferably, the steps for obtaining the vulnerable line risk contribution list are as follows:

[0015] Based on the network-wide cascading risk quantification index, a pre-acquired safety threshold value is set, and the network-wide cascading risk quantification index is compared with the safety threshold value to determine whether the network-wide cascading risk quantification index exceeds the safety threshold value, thereby generating a risk over-limit judgment result.

[0016] Based on the risk exceeding the limit judgment result, if the cascaded risk quantification index of the entire network exceeds the safety threshold value, then backtrack to the disconnected line number and the power flow comparison table of each line after the accident, traverse the scenario of each disconnected line and extract the corresponding risk contribution value set, and sum the single-line risk contribution value in each risk contribution value set to form the cumulative risk contribution value of each line fault scenario.

[0017] Based on the cumulative risk contribution value of each line fault scenario, the cumulative risk contribution value is sorted numerically, and the line with the highest cumulative risk contribution value is selected as the initial fault line. The initial fault line and its corresponding cumulative risk contribution value are recorded to form a list of vulnerable line risk contributions.

[0018] Preferably, the step of obtaining the microgrid group's preventive power flow guidance instruction set is as follows:

[0019] Based on the risk contribution list of vulnerable lines, the cumulative risk contribution value and line number of each line in the risk contribution list are extracted one by one. The line numbers are sorted according to the cumulative risk contribution value. In combination with the microgrid distribution of each line, the power injection or power absorption adjustment value required for the corresponding microgrid is determined one by one, and a set of microgrid group power adjustment values ​​is generated.

[0020] Based on the power adjustment value set of the microgrid group, the power injection or power absorption adjustment value corresponding to each microgrid unit is mapped one by one to an executable power flow diversion command. The power flow diversion commands are then assigned to the corresponding microgrid units one by one with the microgrid unit number as the index, forming a microgrid group preventive power flow diversion command set.

[0021] Preferably, the steps for obtaining the cluster-wide out-of-limit state information set are as follows:

[0022] Based on the execution status of the microgrid cluster preventive power flow guidance instruction set, each microgrid unit continuously collects the current microgrid unit's frequency deviation value, the energy storage device's state of charge value, and the operating status change value of neighboring microgrid units through a local monitoring device, and compares them one by one with their corresponding preset safe operation threshold values. The value entries that exceed the preset safe operation threshold values ​​and the microgrid unit number are recorded to form a cluster-wide over-limit status information set.

[0023] Preferably, the step of obtaining the hierarchical event trigger activation signal is as follows:

[0024] Based on the cluster-wide over-limit status information set and the command update amount issued by the cluster coordinator, the frequency deviation value, energy storage state of charge change value, and neighboring cluster power exchange change value of each microgrid unit are read one by one. At the same time, the frequency adjustment target value, energy storage charging and discharging plan value, and power exchange expectation value of the microgrid unit in the same time period are matched and extracted. The absolute value of the difference between the state deviation and the command target value is calculated respectively, and compared with the corresponding frequency deviation threshold, state of charge threshold, and power exchange deviation threshold in turn. The microgrid unit identifier and deviation type of any difference absolute value exceeding the set threshold are selected, and a hierarchical event trigger activation signal is generated.

[0025] Preferably, the step of obtaining the cluster-level optimal control sequence is as follows:

[0026] Based on the activation signal triggered by the hierarchical event, calculate the individual dynamic adjustment urgency index for each microgrid;

[0027] Based on the individual dynamic adjustment urgency index of each microgrid unit, they are sorted from high to low, and the state characteristics of microgrid units with index values ​​in the top 30% range are retrieved in turn. Adjustable power range, response delay and current energy storage state of charge are extracted. Combined with the load change and expected adjustment effect of various execution actions, the optimal set of control actions is screened and assembled into a continuous adjustment sequence to generate a cluster-level optimal control sequence.

[0028] Preferably, the step of obtaining the microgrid dynamic frequency support scheduling scheme is as follows:

[0029] Based on the cluster-level optimal control sequence, the planned power and planned time period of the energy storage device of each microgrid unit in the cluster-level optimal control sequence are extracted one by one. Combined with the current actual state of charge value of the corresponding microgrid unit and the upper and lower limits of the allowed state of charge interval, the predicted state of charge after the implementation of the planned power is compared one by one to see if it exceeds the allowed interval. Microgrid units whose state of charge meets the interval constraint conditions are selected, and the control sequence after the energy storage state of charge constraint verification is generated.

[0030] Based on the control sequence after the energy storage state of charge constraint verification, the planned power adjustment range of the generator set in each microgrid unit in the sequence is obtained one by one. Combining the allowable range value of the ramp rate of the corresponding generator set with the planned adjustment time interval value, the actual ramp rate of the generator set is calculated. The actual ramp rate is compared one by one to see if it exceeds the allowable range. The adjustment items of the microgrid unit that do not exceed the ramp rate constraint are retained, and the control sequence after the generator ramp rate constraint verification is generated.

[0031] Based on the control sequence after the generator ramp rate constraint verification, the predicted line transmission power value after each microgrid unit in the sequence performs frequency support action is retrieved one by one. The maximum allowable transmission power of the corresponding line is extracted, and the percentage of the transmission power value to the maximum allowable value is calculated. If the percentage exceeds 100%, the power allocation value of the corresponding microgrid unit is adjusted one by one to meet the line transmission capacity constraint condition, and the frequency support power adjustment and allocation under the constraint condition is completed, generating a microgrid dynamic frequency support scheduling scheme.

[0032] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0033] This invention acquires real-time power grid topology and line power flow data, performs virtual disconnection verification on key transmission lines one by one, and realizes simulation and analysis of line fault scenarios. Based on this, it establishes a network-wide cascaded risk quantification index and identifies the initial fault line with the greatest threat to system security by comparing it with a safety threshold in real time. It generates a risk contribution list of vulnerable lines, enabling the location and early prevention of high-risk lines. Furthermore, it directly calculates the power adjustment required by the microgrid group based on the risk contribution list, forming a forward-looking power flow guidance instruction set. Each microgrid unit judges the difference between the actual operating state and the instruction update amount, accurately triggers hierarchical event signals, and quickly starts collaborative optimization calculation to form a cluster-level optimal control sequence adapted to the current system operating state and load demand. Then, it performs real-time verification and allocation of the control sequence under global constraints such as energy storage charge state, generator ramp rate, and line transmission capacity, and establishes a microgrid dynamic frequency support scheduling scheme suitable for multi-constraint conditions. This improves the dynamic response capability of the new energy power system to local disturbances, reduces the cascading risk of single-point faults spreading to the whole network, ensures the stability of power grid operation, and improves the efficiency of new energy absorption. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of the steps of the present invention. Detailed Implementation

[0035] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0036] Please see Figure 1 This invention provides a technical solution: an optimal decentralized coordinated control method for microgrids in a new energy power system, comprising the following steps:

[0037] Based on the power grid topology and real-time line power flow data acquired online, the key transmission lines are virtually disconnected one by one to establish a network-wide cascaded risk quantification index.

[0038] Based on the network-wide cascaded risk quantification index, a numerical comparison is made with the safety threshold. When the network-wide cascaded risk quantification index exceeds the threshold, the initial fault line with the highest contribution to the network-wide cascaded risk quantification index is identified, a vulnerable line risk contribution list is generated, and the power injection or absorption adjustment amount executed by the microgrid group is calculated and determined based on the vulnerable line risk contribution list, and the microgrid group's preventive power flow guidance instruction set is obtained.

[0039] During the execution of the microgrid group's preventive power flow guidance instruction set, each microgrid unit continuously monitors and generates a set of in-cluster limit-breaking status information. The in-cluster limit-breaking status information set and the instruction update amount issued by the cluster coordinator are compared with a preset threshold to generate a hierarchical event trigger activation signal.

[0040] Based on the hierarchical event-triggered activation signal, each cluster coordinator initiates collaborative optimization calculations to generate the cluster-level optimal control sequence. The cluster-level optimal control sequence is then checked and allocated to establish a microgrid dynamic frequency support scheduling scheme.

[0041] The steps to obtain the network-wide cascading risk quantification indicators are as follows:

[0042] Based on the power grid topology information collected online and the power flow values ​​of each transmission line in the current state, a simulation of disconnection operation is performed on any one line in the set of key transmission lines. The power flow change values ​​of the remaining lines are recorded, and the total power flow amplitude under the accident scenario is calculated by combining the power flow values ​​before disconnection. A table of disconnected line numbers and power flow comparison of each line after the accident is generated.

[0043] Based on the disconnected line number and the power flow comparison table of each line after the accident, the power flow increment and initial load power before and after the accident are extracted, and the load rate of each line in the current disconnection scenario is calculated. At the same time, the proportion of the initial power flow in the total power flow of the entire network is extracted, and a set of risk contribution values ​​of all lines in the disconnection scenario is generated.

[0044] Based on the risk contribution value set of all lines in the disconnection scenario, the cascading risk quantification index of the entire network is calculated using the following formula:

[0045] ;

[0046] in, This represents a quantitative indicator of cascading risks across the entire network. This represents the set of critical transmission lines. Indicates the critical transmission line number that was disconnected. This indicates the route numbers of the remaining routes participating in the evaluation. This represents the set of all transmission lines. Indicates the line Current flow value before disconnection Indicates the line On the line The power flow increment after disconnection Indicates the line Maximum tidal carrying capacity Indicates the risk load threshold. Indicates the overload nonlinear amplification index. Indicates the first The absolute value of the power flow before the line is disconnected is used to calculate the line's power flow. Its relative importance in the overall online trend.

[0047] Specifically, based on the online-collected power grid topology information and the current power flow values ​​of each transmission line, the first step is to determine the set of critical transmission lines. This set is not randomly selected but rather determined through a quantitative evaluation system. Specifically, the system retrieves the entire network's topology every 5 seconds from the SCADA and WAMS systems, including data on all nodes, branches, transformers, and generators, as well as the real-time active and reactive power flow values ​​of each transmission line. For each line, its criticality is calculated based on three dimensions: line load factor, network centrality, and historical failure rate. The line load factor is obtained by dividing the real-time power flow value by its static thermal stability limit capacity. Network centrality is calculated using the betweenness centrality algorithm, which counts the number of times each line appears on the shortest path between all pairs of nodes in the power grid. Historical failure rate is extracted from the historical failure record database of the operation and maintenance department, counting the number of unplanned outages per unit length of the line in the past three years. Weights are assigned to these three dimensions, for example, load rate weight is 0.5, network centrality weight is 0.3, and historical failure rate is 0.2. The normalized values ​​of the three dimensions of each line (scaled to between 0 and 1 using the min-max normalization method) are multiplied by their corresponding weights and summed to obtain the criticality score of each line. A criticality threshold is set, for example, 0.75. All lines with scores higher than this threshold are included in the critical transmission line set. Subsequently, each line in the critical transmission line set (e.g., line m) is subjected to a virtual disconnection operation, that is, the connection of the line is disconnected in the simulation model, and the power flow distribution of the entire power grid is recalculated. The simulation program will output a new power flow value that includes all other lines (e.g., line i) after this virtual incident. The dataset is obtained by comparing the post-accident power flow values ​​with the pre-accident initial power flow values. Subtracting these values ​​yields the power flow change values ​​for each line. At the same time, the power flow values ​​before the disconnection and the total power flow amplitude after the accident (the sum of the absolute values ​​of the power flow of all lines) are recorded together. Finally, a detailed reference table is generated for each virtual disconnected critical line. The table uses the disconnected line number (m) as the primary key and lists in detail the numbers, initial power flow values, accident background power flow values ​​and power flow change values ​​of all other lines (i) in the network, generating a reference table of disconnected line numbers and power flow of each line after the accident.

[0048] Based on the disconnected line number and the power flow comparison table of each line after the accident, the system performs an in-depth analysis of each virtual disconnection scenario and calculates the risk contribution factor for each undisconnected line in each scenario. Specifically, the system first traverses each disconnection scenario generated in the previous step (identified by the disconnected line m) in the comparison table. Within each scenario, it then processes all the remaining undisconnected lines i one by one. For each line i, the system extracts its power flow value before the accident from the comparison table. ) and the power flow value after the accident ( At the same time, the maximum allowable transmission power of the line is retrieved from the power grid basic parameter database. The load factor is calculated by dividing the absolute value of the power flow after the accident by the maximum allowable transmission power. To quantify the stress level of line i under an accident, and at the same time, to assess the systemic importance of line i in the entire power grid, the system calculates the proportion of its initial power flow in the total power flow of the entire network. This proportion is calculated by dividing the absolute value of the initial power flow of line i by the sum of the absolute values ​​of the initial power flow of all lines in the entire network (all lines j in set N). This ratio represents the power transmission responsibility weight undertaken by the line during normal operation. Subsequently, the importance weight of the line is combined with the overload risk level of the line under an accident to form the risk contribution value of the line under the specific disconnection scenario. This risk contribution value integrates the initial importance of the line and the severity of the state after the accident, and is the basic unit for constituting the cascaded risk index of the entire network. The system will calculate such a risk contribution value for all other lines i under the current disconnection scenario m, and collect these values ​​to form a risk contribution value list corresponding to scenario m. Finally, after traversing all disconnection scenarios of critical lines, the system will obtain multiple risk contribution value lists, each of which corresponds to an initial fault. These lists together constitute the risk contribution value set of all relevant lines in the entire network under all preset severe accident scenarios.

[0049] formula: The advantage of the formula is that it works through the outer layer. The operation focuses on the most severe single failure scenario under the N-1 criterion, making risk assessment more targeted and able to identify the greatest threats to system security. Secondly, the internal summation term... The formula cumulatively accounts for the network-wide impact of a single fault, rather than focusing solely on the single line with the most severe overload, thus enabling a more accurate capture of potential trends in fault propagation and cascading tripping. Furthermore, the formula incorporates power flow importance weights. This allows core lines carrying high power to occupy a higher proportion in risk assessment, aligning with the physical realities of power grid operation. Finally, a nonlinear overload penalty term is introduced. By risk load threshold Interference from lightly loaded and normally loaded lines is filtered out, and nonlinear amplification is applied. This dramatically amplifies the risk contribution of heavily overloaded lines, improving the accuracy of risk assessment and the effectiveness of early warning.

[0050] This represents the set of critical transmission lines. This set is determined by comprehensively evaluating all transmission lines in the power grid. The evaluation criteria combine the static and dynamic attributes of the lines, and the specific calculation formula is as follows: ,in The total score for the criticality of the route. , , These are the normalized line load rate, betweenness centrality, and historical failure rate, respectively. Weighting coefficients. , , The setting is determined based on power grid operation experience and dispatch strategies. For example, for a power grid where stability is the primary objective, it can be set to... Line load rates are obtained in real time from the SCADA system, betweenness centrality is calculated through graph theory analysis of the current power grid topology, and historical failure rates are retrieved from the equipment management system database. All lines with scores higher than a preset threshold (e.g., a total score greater than 0.8) are included in the set. For example, if a power grid has 100 lines, and after calculation using the above method, the scores for lines L12, L35, and L78 are 0.85, 0.91, and 0.82 respectively, all exceeding 0.8. .

[0051] This represents the set of all transmission lines. This set is fundamental information about the power grid topology and is provided in real-time by the power grid's Energy Management System (EMS) or Supervisory Control and Data Acquisition (SCADA) system. It is a list containing unique identifiers for all transmission lines in the power grid. This set is dynamically updated when new lines are put into operation or old lines are decommissioned. This will change accordingly. At a specific moment during risk calculation, the system locks the current power grid topology and generates a static set of lines for this calculation. For example, in a regional power grid, there are currently 200 operating transmission lines, numbered from L001 to L200, then the set... That is .

[0052] Indicates the line The power flow value before disconnection. This data is a core parameter for power grid condition estimation or real-time measurement, acquired with high precision and high frequency through intelligent electronic devices (IEDs) or phasor measurement units (PMUs) deployed in substations. The acquired data is uploaded to the SCADA system database in real time via a communication network. Power flow value typically refers to active power, measured in megawatts (MW). Before performing risk assessment calculations, the system extracts a snapshot of the power flow of all lines in the entire network from the SCADA database for the previous sampling period (e.g., the previous 5 seconds). For example, during risk assessment, the system queries the real-time active power flow of line L45 at the calculation time. .

[0053] Indicates the line On the line The power flow increment after a disconnection. This value cannot be measured directly and must be obtained through power system simulation calculations. The specific process is as follows: First, based on real-time grid topology and status data obtained from SCADA, a simulation model that perfectly matches the actual grid is built in power system analysis software (such as PSASP, BPA, or PowerFactory). Then, the critical circuit is simulated in the model. The network is disconnected (N-1 fault), and a power flow calculation program is run to obtain the data for all remaining lines in the network after the fault. The new trend value is denoted as The power flow increment is the difference between the simulated background power flow and the initial power flow. For example, the initial power flow of line L67. The power is 200 MW. When the simulated critical line L12 is disconnected, the simulation calculation yields a new power flow at L67. If it is 280 MW, then .

[0054] Indicates the line The maximum power flow carrying capacity. This is a static design parameter determined by the physical properties of the line, mainly including conductor type, cross-sectional area, and meteorological conditions of the installation environment (such as ambient temperature and wind speed). This value is usually divided into normal transmission limit, short-term overload limit, and accident overload limit. In this application, the long-term thermal stability limit of the line is used. These data are pre-stored in the power company's asset management database or line parameter database and are directly called during calculation. For example, for a 220kV double-split LGJ-400 / 35 type conductor, according to the design specifications and regional meteorological conditions, its long-term thermal stability limit in summer can be found in the database. .

[0055] This represents the risk load threshold. This is a crucial engineering setting parameter used to define when a line enters a "risk" state requiring attention, avoiding unnecessary calculations for a large number of normally operating lines. Its setting is primarily based on industry safety regulations and power system operating guidelines. For example, after an N-1 fault, the line load rate should not exceed 100% of its emergency overload capacity. To allow for a certain safety margin and provide early warning, it can be... The threshold is set between 85% and 95% of the emergency overload limit. The setting process involves analyzing N-1 incident data from the power grid over the past five years, statistically analyzing the average overload level that causes other lines to trip, and for example, if most cascading trips occur when the load rate of the affected lines reaches 90% or higher, then the threshold can be set to... This value means that a line's risk is only included in the overall risk index when its load rate exceeds 90% after an incident.

[0056] This represents the overload nonlinear amplification index. This index is used to simulate the nonlinear physical characteristic where the probability of line faults increases sharply with the degree of overload. Its value reflects the severity of the penalty for the overload. A lower value... A value (e.g., 1) represents a linear penalty, while a higher value (e.g., 2 or 3) indicates an exponential increase in the penalty for severe overload. This value can be determined by statistically fitting historical overload tripping data. Specifically, a large number of data points on line overload and whether a line eventually trips are collected. Using the line overload rate (the portion exceeding 100%) as the independent variable and the tripping probability as the dependent variable, a nonlinear regression analysis is performed to fit a curve, thereby determining the penalty. The optimal value for is determined empirically in engineering when sufficient statistical data is unavailable. .

[0057] Calculation process:

[0058] A power grid has 4 lines After criticality assessment, the set of critical paths was determined as follows: .

[0059] The power grid parameters are as follows:

[0060] Line L1: , ;

[0061] Line L2: , ;

[0062] Line L3: , ;

[0063] Line L4: , ;

[0064] The selected parameters are: , .

[0065] Calculate the total power flow of the entire network

[0066] ;

[0067] Perform N-1 simulation

[0068] Simulate the disconnection of critical path L1 (i.e.) The power flow increment is calculated using simulation software.

[0069] (L2 current increase);

[0070] (L3 current increase);

[0071] (The L4 current has decreased, which may indicate a current reversal or redistribution).

[0072] Calculate the total risk contribution under the L1 disconnection scenario;

[0073] because There is only L1, so This refers to the risk value generated when L1 is disconnected. It is necessary to calculate and sum the risk contributions of L2, L3, and L4, excluding L1.

[0074] Calculate the risk contribution value of line L2 :

[0075] Behind-the-scenes trends of the accident: ;

[0076] Post-accident load rate: ;

[0077] Overload penalty items: ;

[0078] Importance weight: ;

[0079] Risk contribution value: ;

[0080] Calculate the risk contribution value of line L3 :

[0081] Behind-the-scenes trends of the accident: ;

[0082] Post-accident load rate: ;

[0083] because The route did not reach the risk threshold.

[0084] Overload penalty items: ;

[0085] Risk contribution value: ;

[0086] Calculate the risk contribution value of line L4 :

[0087] Behind-the-scenes trends of the accident: ;

[0088] Post-accident load rate: ;

[0089] because The route did not reach the risk threshold.

[0090] Overload penalty items: ;

[0091] Risk contribution value: ;

[0092] Summation:

[0093] ;

[0094] Calculate the final network-wide cascading risk quantification index ;

[0095] Because of the critical path set There is only L1, so:

[0096] ;

[0097] The result indicates that, under the current power grid conditions, the quantifiable risk index for the network-wide cascading failure caused by a line break in critical line L1 is 0.0016875. If the calculated... If the value is less than this threshold, the system is considered to be in a safe state; if... A value greater than 0.01 indicates a high risk of cascading failures in the system, requiring immediate implementation of preventative control measures. In this example, 0.0016875 is significantly less than 0.01, indicating that the system is quite robust in handling L1 disconnection failures, and the risk is manageable. This indicator identifies that the risk primarily stems from the overload stress on L2 caused by the disconnection of L1.

[0098] The steps to obtain the list of vulnerable line risk contributions are as follows:

[0099] Based on the network-wide cascading risk quantification index, a pre-acquired security threshold value is set, and the network-wide cascading risk quantification index is compared with the security threshold value to determine whether the network-wide cascading risk quantification index exceeds the security threshold value, and a risk over-limit judgment result is generated.

[0100] Based on the risk exceeding the limit judgment result, if the cascaded risk quantification index of the entire network exceeds the safety threshold value, then backtrack to the disconnected line number and the power flow comparison table of each line after the accident, traverse the scenario of each disconnected line and extract the corresponding risk contribution value set, and sum up the single-line risk contribution value in each risk contribution value set to form the cumulative risk contribution value of each line fault scenario.

[0101] Based on the cumulative risk contribution value of each line fault scenario, the cumulative risk contribution value is sorted numerically, and the line with the highest cumulative risk contribution value is selected as the initial fault line. The initial fault line and its corresponding cumulative risk contribution value are recorded to form a list of vulnerable line risk contributions.

[0102] Specifically, based on the network-wide cascaded risk quantification index, the system first calls a preset safety threshold value to determine the risk level. This safety threshold is set based on statistical analysis of historical power grid operation data. Specifically, the system retrieves a historical dataset of the network-wide cascaded risk quantification index, calculated every 5 minutes over the past year, forming a long-term series containing approximately 105,120 data points. After removing extreme outliers caused by data collection errors, the mean and standard deviation of this dataset are calculated. The safety threshold is set using a dynamic grading strategy, divided into three levels: "Attention," "Warning," and "Severe." The calculation method is as follows: the "Attention" level threshold is set to the historical data mean plus 1.5 times the standard deviation; the "Warning" level threshold is set to the historical data mean plus 3 times the standard deviation; and the "Severe" level threshold is set to... The threshold for each level is set as the historical data mean plus 5 times the standard deviation. For example, if the historical risk indicator mean is 0.005 and the standard deviation is 0.002, then the "attention" threshold is 0.005 + 1.5 × 0.002 = 0.008, and the "warning" threshold is 0.005 + 3 × 0.002 = 0.011. The "safety threshold value" used in this step is the "warning" level threshold of 0.011. The system directly compares the currently calculated cascading risk quantification indicator for the entire network, for example, 0.015, with this safety threshold value of 0.011. Since 0.015 is greater than 0.011, the comparison logic returns "true". Based on this, the system determines that the current power grid operation status has exceeded the preset safety range and there is a significant risk of cascading failure, generating a risk over-limit judgment result.

[0103] Based on the risk exceeding the limit judgment result, when the judgment result is that the risk exceeds the limit, the system immediately starts the risk tracing analysis program. This program first backtracks and retrieves the intermediate data cache generated during the calculation of the cascaded risk quantification index of the entire network, namely the "Disconnected Line Number and Power Flow Comparison Table of Each Line after the Accident". This comparison table records in detail the power flow response of all other lines in the entire network after a virtual disconnection operation is performed on each critical transmission line. The program uses the critical line number as an index to start traversing each virtual fault scenario in the comparison table. For each disconnected line m, the program extracts the initial power flow value of all other lines i in that scenario. ), trend increment ( The system recalculates the single-line risk contribution value of each line i in this scenario. The calculation process strictly follows the internal term in the formula of the whole network cascade risk quantification index, that is, combining the importance weight of the line (the proportion of its initial power flow to the total power flow of the whole network) and the nonlinear overload penalty degree. The single-line risk contribution values ​​of all non-faulty lines (i≠m) in this scenario are arithmetically summed to obtain a total value. This total value is the risk intensity of the simulated line m when it fails, which is defined as the risk contribution accumulation value of this scenario. The system stores this accumulation value and the corresponding disconnected line number m as a data pair. For example, when analyzing the disconnection scenario of the critical line L35, the risk contribution values ​​of all other lines are calculated and accumulated to obtain a total of 0.028. Then (L35, 0.028) is recorded. The program repeats this process until all the disconnection scenarios of the critical transmission lines are traversed, forming the risk contribution accumulation value of each line fault scenario.

[0104] Based on the cumulative risk contribution value of each line fault scenario, the system processes this set containing multiple data pairs (critical line number, cumulative risk contribution value). First, a standard descending sorting algorithm is executed to sort all critical lines from largest to smallest according to the cumulative risk contribution value in each data pair. After sorting, the system does not select all lines, but selects the most critical fault source according to a preset "top-level risk source screening rule". This rule is defined as selecting a fixed number of lines at the top of the sorted list, such as selecting the top 5 lines. This number (5) is preset according to the grid scale and the dispatchability of prevention and control resources. For a 500 kV backbone network, this number is usually set between 3 and 5. To concentrate resources on addressing the most significant risks, for example, the sorted list is [(L35, 0.028), (L72, 0.025), (L18, 0.019), (L91, 0.011), (L46, 0.009), ...]. Following the rule of selecting the top 3, the system will select L35, L72, and L18 as the initial fault lines that pose the greatest threat to power grid security. Subsequently, the system will extract the numbers of these three lines and their corresponding cumulative risk contribution values, and organize them into a structured data list. Each row of the list contains two fields: "Initial Fault Line Number" and "Cumulative Risk Contribution Value". This newly generated list is the final output result, forming a list of risk contributions for vulnerable lines.

[0105] The steps for obtaining the microgrid group's preventative power flow control instruction set are as follows:

[0106] Based on the list of risk contributions of vulnerable lines, the cumulative risk contribution value and line number of each line in the list are extracted one by one. The line numbers are sorted according to the cumulative risk contribution value. In combination with the microgrid distribution of each line, the power injection or power absorption adjustment value required for the corresponding microgrid is determined one by one, and a set of microgrid group power adjustment values ​​is generated.

[0107] Based on the power adjustment value set of the microgrid group, the power injection or power absorption adjustment value corresponding to each microgrid unit is mapped one by one to an executable power flow diversion command. The power flow diversion command is then assigned to the corresponding microgrid unit one by one with the microgrid unit number as the index, forming a preventive power flow diversion command set for the microgrid group.

[0108] Specifically, based on the list of vulnerable lines' risk contributions, the system first processes the items in the list, which is sorted from highest to lowest by cumulative risk contribution value. For example, it includes (L35, 0.028), (L72, 0.025), and (L18, 0.019). Next, for each vulnerable line in the list, the system initiates a power adjustment demand calculation process. Taking the highest-risk line L35 as an example, the system first queries a pre-set "Grid Topology and Microgrid Association Database" for the microgrid unit with the closest electrical distance to line L35. This electrical distance is quantified by calculating the equivalent impedance from the nodes at both ends of the line to the grid connection point of each microgrid. All microgrids with an equivalent impedance less than 0.05 per-unit are considered strongly correlated microgrids. For example, the query result would be microgrids MG1, MG2, and MG3. Subsequently, the system needs to calculate an overall power adjustment target value, which is directly related to the cumulative risk contribution value of the line. The calculation formula is as follows: ,in It is the cumulative value of risk contribution. This is a risk power conversion coefficient, which is preset by the power grid dispatch center based on the current system's reserve capacity and stability margin. For example, if it is set to 500 MW / risk unit, then the total adjustment power required for line L35 is 0.028 × 500 = 14 MW. Next, this 14 MW adjustment needs to be allocated to microgrids MG1, MG2, and MG3. The allocation is based on each microgrid's sensitivity to the power flow of line L35 and its own available adjustment capacity. The system calculates the sensitivity factor of each microgrid's power injection to changes in the power flow of line L35 using disturbance analysis, and obtains the adjustable power margins from each microgrid's local controller in real time. Finally, the power adjustment amount for microgrid j is calculated as follows: ,in It is the power flow sensitivity of microgrid j to line L35. The available adjustable power is calculated sequentially using this method to determine the power adjustment values ​​for MG1, MG2, and MG3, for example, +6.5 MW, +4.5 MW, and +3.0 MW (positive values ​​indicate power injection). This process is repeated for all vulnerable lines in the list, and then all calculated microgrid power adjustment values ​​are summarized to generate a set of microgrid group power adjustment values.

[0109] Based on the power adjustment value set of the microgrid group, the system initiates an instruction generation and distribution program. This program reads each entry in the set, for example (MG1, +6.5 MW), and converts it from a pure numerical value into a structured instruction containing detailed execution parameters. This mapping process first parses the sign of the power adjustment value; a positive sign is mapped to a "power injection" instruction type, and a negative sign is mapped to "power absorption." The absolute value is the target amplitude of the power adjustment. Next, the instruction must include the adjustment rate, i.e., the speed of power change. This rate is not arbitrarily set but determined based on the microgrid's own dynamic capabilities. The system queries the pre-stored microgrid parameter table to obtain the maximum allowable power ramp rate of microgrid MG1, for example, 2 MW / min. To avoid excessive equipment wear and maintain a certain margin, the ramp rate in the command is set to 70% of the maximum value, i.e., 1.4 MW / min. Furthermore, the command needs to specify the start and end times or duration. In preventative control scenarios, a standard control cycle is typically set, such as 15 minutes. This means that after the command is issued, the microgrid needs to reach the target power within approximately 4.6 minutes (6.5 MW divided by 1.4 MW / min) and maintain this output until the cycle ends or a new command is received. Finally, the system generates a unique global command ID and a timestamp for this command, assembling them into a complete power flow control command. Its format can be a JSON object: {"Command ID": "CMD-20231027-001"}. The system repeats this mapping process for all entries in the microgrid group's power adjustment value set, including: “Microgrid Unit Number”: “MG1”, “Command Type”: “Power Injection”, “Power Adjustment Value”: “6.5 MW”, “Execution Ramp Rate”: “1.4 MW / min”, “Duration”: “15 minutes”, and “Timestamp”: “2023-10-27T10:30:00Z”. All generated structured commands are packaged and distributed to the local controller of each corresponding microgrid unit via an encrypted industrial communication protocol, such as a GOOSE message based on IEC 61850, with the microgrid unit number as the target address, forming a microgrid group preventative power flow guidance command set.

[0110] The steps for obtaining the cluster-wide out-of-bounds state information set are as follows:

[0111] Based on the execution status of the microgrid cluster's preventive power flow guidance instruction set, each microgrid unit continuously collects the current microgrid unit's frequency deviation value, the energy storage device's state of charge value, and the operating status change value of neighboring microgrid units through local monitoring devices, and compares them one by one with their corresponding preset safe operation threshold values. The value entries that exceed the preset safe operation threshold values ​​and the microgrid unit number are recorded to form a cluster-wide over-limit status information set.

[0112] Specifically, based on the execution status of the microgrid cluster's preventative power flow guidance command set, the local controller of each microgrid unit continuously collects key operating parameters through its internal monitoring network. This process is performed at a high frequency. For example, the frequency deviation value is obtained by sampling 10 times per second using a local frequency meter synchronized with GPS, and the difference from the grid's nominal frequency (e.g., 50 Hz) is calculated. The state of charge value of the energy storage device is reported once per second by the battery management system (BMS), which directly reflects the energy level of the energy storage unit. The operating status changes of neighboring microgrid units are obtained by monitoring the bidirectional power flow on the inter-microgrid interconnects, and the change value is obtained by calculating the difference between the current power reading and the reading one second ago. The controller continuously compares these collected real-time data streams with a series of preset safe operating thresholds. These thresholds are set with clear criteria. The safe threshold for frequency deviation is set at ±0.2 Hz, which refers to the definition of first and second level frequency deviations in the grid operation regulations. The safe operating threshold for the energy storage state of charge is a range, such as [20%, 90%]. This range is provided by the energy storage battery manufacturer. Operating outside this range will accelerate battery aging or cause safety issues. The threshold for changes in the operating status of neighboring microgrids is specifically defined as the threshold for the change rate of tie-line power exchange, set at 1 MW per second. This value is the upper limit of power fluctuation that will not cause oscillations, determined based on simulation analysis of the dynamic stability of the microgrid cluster. When any monitored value touches or exceeds its corresponding threshold boundary, for example, when the frequency deviation of microgrid MG2 reaches -0.21 Hz, or its energy storage state of charge drops to 19.8%, the controller will immediately generate an over-limit record. This record includes the microgrid cell number, the timestamp of the event, the name of the over-limit parameter (such as "frequency deviation"), the measured value (such as -0.21 Hz), and the corresponding threshold (±0.2 Hz). All multiple over-limit records generated by the microgrid within the same evaluation period are aggregated to form an intra-cluster over-limit state information set.

[0113] The steps for obtaining the activation signal triggered by the layered event are as follows:

[0114] Based on the cluster-wide over-limit state information set and the command update amount issued by the cluster coordinator, the frequency deviation value, energy storage state of charge change value, and neighboring cluster power exchange change value of each microgrid unit are read one by one. At the same time, the frequency adjustment target value, energy storage charging and discharging plan value, and power exchange expectation value of the microgrid unit in the same time period are matched and extracted. The absolute value of the difference between the state deviation and the command target value is calculated respectively, and then compared with the corresponding frequency deviation threshold, state of charge threshold, and power exchange deviation threshold in turn. Microgrid unit identifiers and deviation types of any difference absolute value exceeding the set threshold are selected, and a hierarchical event trigger activation signal is generated.

[0115] Specifically, based on the cluster-wide out-of-limit state information set and the command update amount issued by the cluster coordinator, the system initiates an event-triggered decision logic. This logic first parses each record in the cluster-wide out-of-limit state information set, for example, a record is (MG2, timestamp T1, frequency deviation, -0.21 Hz). Simultaneously, the system matches the control target of MG2 within the similar time period T1 from the latest command update amount issued by the cluster coordinator. This command update amount is a set containing the refined targets of all microgrid units, for example, including (MG2, frequency regulation target value, -0.05 Hz, energy storage). The planned charge / discharge value is -1.2 MW, and the expected power exchange value is +0.5 MW. The system then calculates the deviation between the measured state value and the target value. For frequency, the deviation is |-0.21 - (-0.05)| = 0.16 Hz. For the energy storage state of charge, the deviation between its change value (e.g., the monitored current state of charge change rate is -1.5 MW, i.e., the discharge rate) and the planned value (-1.2 MW) is |-1.5 - (-1.2)| = 0.3 MW. For power exchange, the actual change value (e.g., +0.2 MW) and the expected value (+0.5 MW) are calculated. The deviation is |+0.2-(+0.5)| = 0.3 MW. The system then compares the absolute values ​​of these calculated differences with their corresponding event trigger thresholds. These thresholds are preset, designed to sensitively detect significant deviations while avoiding frequent triggering due to minor disturbances. The frequency deviation threshold is set at 0.1 Hz, which is five times the dead zone of the primary frequency regulation of the power grid. The state-of-charge deviation threshold is set at 5% of the rated power of the energy storage system; for example, for a 5 MW energy storage system, the threshold is 0.25 MW. The power exchange deviation threshold is set at 2% of the maximum capacity of the tie line. For example, for a 20 MW tie line, the threshold is 0.4 MW. In this case, the frequency deviation of MG2 is 0.16 Hz, which exceeds the threshold of 0.1 Hz, and the energy storage power deviation is 0.3 MW, which exceeds the threshold of 0.25 MW. However, the power exchange deviation is 0.3 MW, which does not exceed the threshold of 0.4 MW. Since at least one deviation exceeds the limit, the system determines that MG2 has triggered an event and records its identifier "MG2" as well as the triggered deviation types "frequency deviation" and "energy storage power deviation". The system summarizes the microgrid identifiers and deviation types of all such triggered events and generates a hierarchical event trigger activation signal.

[0116] The steps for obtaining the cluster-level optimal control sequence are as follows:

[0117] Based on the activation signal triggered by the stratified event, the individual dynamic adjustment urgency index of each microgrid is calculated using the following formula:

[0118] ;

[0119] in, Indicates the first Individual dynamic adjustment urgency index of each microgrid unit , , These represent the basic weights for frequency deviation, energy storage power deviation, and power exchange deviation, respectively. The coupling factor representing the frequency response and the deviation from the state of charge. Indicates the first Prediction frequency deviation of individual microgrid cells This is the frequency reference deviation value. Indicates the first The difference between the energy storage power dispatched by a microgrid and the planned power. This indicates the maximum energy storage capacity that the microgrid can utilize. Indicates the first Predicted power exchange deviation between a microgrid and its neighboring clusters This indicates the maximum switching capacity of the micronetwork. Indicates the first Current state of charge of microgrid energy storage , , These are the reference value, maximum value, and minimum value, respectively. The weight of the system imbalance penalty term. This is the frequency-power deviation coupling coefficient;

[0120] Based on the individual dynamic adjustment urgency index of each microgrid unit, they are sorted from high to low, and the state characteristics of microgrid units with index values ​​in the top 30% range are retrieved in turn. Adjustable power range, response delay and current energy storage state of charge are extracted. Combined with the load change and expected adjustment effect of various execution actions, the optimal set of control actions is screened and assembled into a continuous adjustment sequence to generate a cluster-level optimal control sequence.

[0121] Specifically, the formula: The advantage of this formula lies in its nonlinear amplification of the degree of deviation from normal conditions by introducing the square terms of frequency deviation, energy storage power deviation, and tie-line power deviation, thus making the urgency higher for microgrids in worse condition. Secondly, it introduces coupling terms. The system dynamically links the state of charge (SOC) of energy storage with frequency response capability. When the SOC deviates from the ideal reference value, the weight of its frequency deviation is amplified. This gives microgrids with excessively high or low SOC a higher regulation priority when facing frequency fluctuations. Finally, a penalty term similar to Area Control Error (ACE) is introduced. It takes into account frequency deviation and tie-line power deviation, thus suppressing the overall power imbalance of the system that may be caused by a single regulation target.

[0122] , , These represent the basic weights for frequency deviation, energy storage power deviation, and power exchange deviation, respectively. These weights are used to adjust the relative importance of different types of deviations in the final urgency index. Their settings are determined using the Analytic Hierarchy Process (AHP). First, a judgment matrix is ​​constructed, and multiple grid dispatching experts score the pairwise relative importance of these three indicators (using a 1-9 scale). For example, experts generally consider frequency stability to be the primary task, its importance far exceeding that of energy storage or power exchange plan tracking; therefore... right The rating is 5. right The rating is 7; energy storage program tracking is slightly more important than tie-line power stability. right The score is 3. Then, by calculating the largest eigenvalue of the judgment matrix and its corresponding eigenvector, and normalizing the eigenvector, the weights of each indicator are obtained. For example, after calculation, the normalized weight vector is: Therefore, setting , , These weights are dynamically adjusted based on the power grid's operating status (e.g., normal, alarm, emergency). In an emergency, The weight will be further increased.

[0123] This represents the coupling factor between frequency response and state of charge (SOC). This factor quantifies the degree to which the SOC of an energy storage system influences its participation in frequency regulation. When the SOC is in the ideal range (e.g., 40%-60%), the energy storage system responds most flexibly; when the SOC is too high or too low, its continuous charging and discharging capability is limited, and its priority in frequency regulation should be reduced. The value of determines the intensity of this influence. Its setting is based on the analysis of the operating characteristics of the energy storage system. By simulating the continuous frequency regulation power and duration that the energy storage unit can provide under different SOC levels, a nonlinear relationship curve between SOC and frequency regulation capability is plotted. A suitable is selected. The value makes the formula in The term can fit the changing trend of this curve relatively well. Typically, when the SOC deviates from the reference value to the limit (i.e., close to...),... or Its frequency modulation capability may decrease by more than 80%, from which we can deduce... The numerical value. For example, if we require that the weight of the frequency term increases by 100% when the SOC deviates by 50%, then we can solve for the value. An empirical approach is to perform regression analysis on historical data to find the setting that minimizes the overall system adjustment cost. Value. Here, based on simulation and experience, we take... .

[0124] Indicates the first The predicted frequency deviation of each microgrid unit. This parameter is not an instantaneous measurement, but a short-term prediction based on the current frequency change trend, providing more forward-looking control. Its calculation employs a Kalman filter algorithm. This algorithm takes a high-frequency sampled real-time frequency deviation sequence as input to establish a state-space model describing the dynamic changes in frequency. The model comprehensively considers the system's inertia, damping, and random disturbances. Through the Kalman filter's prediction step, the frequency deviation for the next control cycle (e.g., after 5 seconds) can be optimally estimated based on current and historical measurements. For example, in... At time 10:00, the measured frequency deviation was -0.15 Hz, with a rate of change of -0.02 Hz / s. The Kalman filter, combined with the system model, predicted that at 10:00, the frequency deviation was -0.15 Hz. Frequency deviation at time The frequency is -0.24Hz.

[0125] This is the frequency reference deviation value. It's a standardized reference value used to convert absolute frequency deviation into a dimensionless per-unit value, facilitating weighted summation with deviations of different dimensions. Its setting is based on the limits for severe frequency deviations in national or regional power grid operation standards. For example, when the frequency deviation exceeds ±0.5Hz, the system is in an emergency state, requiring emergency control measures such as load shedding. Therefore, it can be... Set to 0.5Hz. This value represents the "full scale" of the frequency deviation, and any actual deviation will be normalized to this value. For example, if... The normalized frequency deviation is pu.

[0126] Indicates the first The difference between the energy storage power dispatched and the planned power of a microgrid. This parameter reflects the degree to which the actual output of the energy storage unit deviates from its predetermined plan. It is calculated by acquiring two data points in real time from the microgrid's local controller: one is the actual output power of the energy storage system. The power is measured directly by the power meter; secondly, it is measured within the current scheduling cycle by the energy storage plan issued by the superior scheduling system. The difference is calculated as follows: For example, the scheduling plan requires the energy storage of microgrid c to discharge at a power of 1.2MW at the current moment (denoted as -1.2MW), but due to participation in a primary frequency regulation, its actual discharge power is 1.5MW (denoted as -1.5MW). .

[0127] This indicates the maximum energy storage capacity that the microgrid can utilize. This is a key design parameter of the energy storage system, representing the maximum active power that its inverter (PCS) can output or absorb. This value is determined by the equipment nameplate parameters and stored in the microgrid's asset database. During calculations, the system retrieves this value directly from the database. For example, if a microgrid is configured with two 1MW / 2MWh energy storage units, its total maximum utilization capacity is [value missing]. This parameter is used to measure the energy storage power deviation. Normalization is performed.

[0128] Indicates the first The predicted power exchange deviation between a microgrid and its neighboring clusters. This parameter measures the degree to which the microgrid's external power exchange behavior deviates from the plan, and is a predicted value. Its calculation combines real-time measurements and short-term load forecasting. First, the actual exchange power between the microgrid and the main grid or neighboring microgrids on the interconnect lines is measured in real time. Then, the planned switching power of the tie line is obtained from the scheduling plan. Meanwhile, an ultra-short-term load forecasting module based on time series analysis (such as the ARIMA model) will predict the net changes in load and distributed generation output within the microgrid during the next control cycle. The predicted power exchange deviation is: For example, if the planned switching power is 5MW to be supplied to the main grid, the actual measured power supply is 4.5MW, and the load is predicted to increase by 0.2MW within the next 5 seconds, then the predicted switching power is 4.5 - 0.2 = 4.3MW, and the prediction deviation is... .

[0129] This indicates the maximum switching capacity of the microgrid. This refers to the maximum allowable transmission power of the tie line connecting the microgrid to the external power grid, determined by the physical parameters of the tie line itself (such as conductor thermal capacity) and the capacity of the transformers at both ends. This is a fixed static parameter recorded in the power grid's equipment ledger. For example, a microgrid is connected to the 110kV main grid via a dedicated 35kV line; the maximum transmission capacity of this line is calculated to be... This parameter is used to measure power exchange deviation. Normalize.

[0130] Indicates the first The current state of charge (SOC) of a microgrid energy storage system. This data is calculated and reported in real time by the system's Battery Management System (BMS). The BMS measures the battery's terminal voltage, current, and temperature, and uses various estimation algorithms, including coulomb integration and open-circuit voltage methods, to determine the percentage of remaining battery capacity relative to the total capacity. For example, in a 2MWh energy storage system, if the BMS reports a current remaining capacity of 1.1MWh, then its... .

[0131] , , These are the reference, maximum, and minimum values ​​for the energy storage state of charge. These are parameters set by the energy storage operation strategy. and These are operating boundaries set to protect the battery from overcharging and over-discharging, and are usually recommended by the battery manufacturer, for example... , . This is the ideal standby SOC level, usually set at the middle position to ensure equal charging and discharging capabilities. The most common setting is... This reference value is also the target SOC for energy storage applications such as energy market arbitrage or peak shaving.

[0132] This is the weight for the system imbalance penalty term. This weight is used to adjust the impact of the ACE term on the overall urgency. When the primary objective of the microgrid cluster is to maintain regional power balance (e.g., in islanded mode or when participating in the ancillary service market), this weight should be set higher. Its value can be determined based on simulation analysis: under different... Under these conditions, a series of disturbance tests are conducted on the microgrid group to evaluate its overall performance in terms of frequency stability and power balance, and a group that achieves the optimal overall performance is selected. Value. Based on experience, in grid-connected mode, this weight is typically lower than the weight for frequency deviation and can be set to [value missing]. .

[0133] This is the frequency-power deviation coupling coefficient. In ACE calculations, this coefficient plays a role similar to the frequency offset coefficient in traditional power systems. It defines how much power adjustment is needed to compensate for a unit frequency deviation, and its unit is MW / Hz. Theoretically, its value is equal to the sum of the microgrid's own load frequency response coefficient and the generator's power-frequency static characteristic. In practice, it can be obtained through online identification or historical data statistics. For example, by analyzing historical data, it is found that for every 0.1Hz decrease in the microgrid's frequency, its total load and the generator's natural response cause it to draw an additional 2MW of power from the main grid. .

[0134] Calculation process:

[0135] Consider microgrid units The microgrid is activated based on the activation signal triggered by the hierarchical event.

[0136] Get parameter values:

[0137] ;

[0138] , ;

[0139] , ;

[0140] , ;

[0141] , , , ;

[0142] Calculate each normalized deviation term:

[0143] Normalized frequency bias: ;

[0144] Normalized energy storage power deviation: ;

[0145] Normalized power exchange bias: ;

[0146] Normalized SOC deviation: ;

[0147] Calculate each urgency component:

[0148] Frequency deviation urgency component:

[0149] ;

[0150] ;

[0151] ;

[0152] ;

[0153] Energy storage power deviation urgency component:

[0154] ;

[0155] Power exchange deviation urgency component:

[0156] ;

[0157] System imbalance penalty component:

[0158] First, it is necessary to Normalization is also performed. .

[0159] ;

[0160] ;

[0161] ;

[0162] ;

[0163] ;

[0164] Calculate the overall individual dynamic adjustment urgency index :

[0165] ;

[0166] This result indicates that microgrid units The individual dynamic adjustment urgency index is 0.2459. This is a comprehensive dimensionless index; the higher the value, the more serious the deviation of the current state of the microgrid from the expectation, and the greater the potential impact on system stability, thus requiring more urgent control and adjustment. In multi-microgrid collaborative optimization scenarios, the cluster coordinator calculates the urgency index of all activated microgrids. Microgrids with high urgency indices (e.g., greater than 0.2) will be prioritized for more regulation tasks, and their internal adjustable resources (such as energy storage and controllable loads) will be utilized preferentially.

[0167] Based on the individual dynamic adjustment urgency index of each microgrid unit, the system executes a sorting and filtering process to determine the microgrid units with priority response. First, the individual dynamic adjustment urgency indices of all activated microgrids are sorted in descending order to form an ordered list. Then, the system selects a portion of the microgrids in this list according to a preset "elite response strategy." This strategy is defined as selecting microgrid units whose urgency indices are in the top 30% range as the core adjustment resource pool. For example, if 10 microgrids are activated, the 3 microgrids with the highest urgency are selected. For these 3 selected microgrids, the system sends a data request command to their local controller through the communication network to retrieve their detailed status characteristic data in sequence. This data includes: adjustable power range, i.e., the maximum upward and downward adjustment power that all controllable resources (such as energy storage, adjustable load, and small generators) can currently provide, for example (-5 MW, +3 MW); and response delay, i.e., the time from receiving the command to the power... The system calculates the average time required for the output to reach 90% of the set value (e.g., 200 milliseconds for energy storage and 10 seconds for diesel generators) and the current energy storage state of charge (e.g., 65%). Then, a rule-based control action filtering module is activated. This module contains a pre-built "control action effect library," which stores the expected adjustment effects (expressed as gain and time delay parameters) of various execution actions (e.g., "rapid discharge of 1 MW of energy storage" and "cutting off 0.5 MW of tertiary loads") on the system's frequency and power balance. The system matches the microgrid's state characteristics with the actions in the effect library, filtering out the most effective and economical combination of control actions in the current state. For example, in a scenario with severely low frequency, the system prioritizes energy storage discharge actions with short response delays and high adjustment power. Finally, the selected optimal control action set is arranged according to the execution order and coordination relationship, assembling into a time-driven continuous adjustment sequence to generate a cluster-level optimal control sequence.

[0168] The steps for obtaining the microgrid dynamic frequency support scheduling scheme are as follows:

[0169] Based on the cluster-level optimal control sequence, the planned power and planned time period of the energy storage device of each microgrid unit in the cluster-level optimal control sequence are extracted one by one. Combined with the current actual state of charge value of the corresponding microgrid unit and the upper and lower limits of the allowable state of charge interval, the predicted state of charge after the implementation of the planned power is compared one by one to see if it exceeds the allowable interval. Microgrid units whose state of charge meets the interval constraint conditions are selected, and the control sequence after the energy storage state of charge constraint verification is generated.

[0170] Based on the control sequence after the energy storage state of charge constraint verification, the planned power adjustment range of the generator set in each microgrid unit in the sequence is obtained one by one. Combining the allowable range of the ramp rate of the corresponding generator set with the planned adjustment time interval, the actual ramp rate of the generator set is calculated. The actual ramp rate is compared one by one to see if it exceeds the allowable range. The adjustment items of the microgrid unit that do not exceed the ramp rate constraint are retained, and the control sequence after the generator ramp rate constraint verification is generated.

[0171] Based on the control sequence after generator ramp rate constraint verification, the predicted line transmission power value after each microgrid unit in the sequence performs frequency support action is retrieved one by one. The maximum allowable transmission power of the corresponding line is extracted, and the percentage of the transmission power value to the maximum allowable value is calculated. If the percentage exceeds 100%, the power allocation value of the corresponding microgrid unit is adjusted one by one to meet the line transmission capacity constraint condition, and the frequency support power adjustment and allocation under the constraint condition is completed, generating a microgrid dynamic frequency support scheduling scheme.

[0172] Specifically, based on the cluster-level optimal control sequence, the system initiates a safety verification program. First, it checks the operational constraints of the energy storage device. This program reads the energy storage scheduling instructions for each microgrid unit in the control sequence line by line. For example, if the instruction for microgrid MG3 is "The energy storage device will continuously discharge at a power of 2 MW for the next 10 minutes," the system immediately obtains the real-time status data of its energy storage device from the local controller of MG3, including the current actual state of charge (SOC), for example, 35%, and the total capacity of the energy storage device, for example, 10 MWh. Simultaneously, the system retrieves the upper and lower limits of the allowed SOC range from the device's configuration database. This range is set by the manufacturer and is typically [20%, 90%]. Next, the system performs a forward simulation calculation to predict the final SOC of the energy storage after executing the instruction. The calculation process is as follows: First, calculate... The total discharge, i.e., 2 MW multiplied by 10 minutes (1 / 6 hour), yields 0.333 MWh. Then, the percentage of this discharge to the total capacity is calculated, i.e., 0.333 divided by 10, resulting in 3.33%. Finally, the current state of charge is subtracted from the discharge percentage, and the predicted final state of charge is 35% - 3.33% = 31.67%. The system compares this predicted value of 31.67% with the allowable range [20%, 90%]. Since 31.67% is within this range, the instruction is determined to meet the state of charge constraint, and the system retains the instruction in the control sequence. If another instruction causes the predicted state of charge to be lower than 20% or higher than 90%, the instruction will be marked as "not met" and removed from the current execution sequence or subject to power correction. After the system has traversed all energy storage scheduling instructions, it generates a control sequence after energy storage state of charge constraint verification.

[0173] Based on the control sequence after the energy storage state of charge constraint verification, the system continues to verify the feasibility of the dispatch instructions involving traditional generator sets in the sequence, focusing on whether the power adjustment rate is within the equipment's allowable range. The system obtains the planned power adjustment range of the generator sets in each microgrid unit in the sequence one by one. For example, the instruction requires the G1 gas turbine in microgrid MG5 to increase its output power from the current 10 MW to 12 MW in the next 5 minutes, that is, the planned power adjustment range is +2 MW. At the same time, the system queries the pre-stored equipment parameter database for the allowable ramp rate range of unit G1. This value is a core performance indicator provided by the generator manufacturer, usually expressed as the percentage of rated power that can be increased or decreased per minute. For example, the rated power of unit G1 is 20 MW, and its maximum allowable positive ramp rate is 5% / minute, that is, it can increase the power by a maximum of 2 MW per minute. Adding 1 MW of output, the system then calculates the actual ramp rate required by the generator set based on the information in the instruction. The calculation method is to divide the planned power adjustment range by the planned adjustment time interval. In this example, the actual ramp rate is 2 MW divided by 5 minutes, which equals 0.4 MW / minute. Then, the system compares the calculated actual ramp rate of 0.4 MW / minute with the maximum allowable ramp rate of 1 MW / minute. Since 0.4 is less than 1, the instruction is determined to meet the ramp rate constraint and is therefore retained in the control sequence. If an instruction requires an actual ramp rate that exceeds the allowable range, such as requiring an increase of 2 MW of power within 1 minute, the instruction will be marked as "infeasible" and removed from the sequence. Alternatively, the system will try to extend its adjustment time to meet the constraint. After the system completes this check on the adjustment items of all generators in the sequence, it generates the control sequence after the generator ramp rate constraint verification.

[0174] Based on the control sequence after generator ramp rate constraint verification, the system enters the final safety verification stage, namely, the transmission capacity verification of transmission lines. The system first uses a power flow calculation engine to perform a rapid "pre-simulation" of the grid state after executing the current control sequence. It retrieves the total power injection or absorption values ​​of each microgrid unit in the sequence after executing all frequency support actions. For example, the total injected power of microgrid MG3 becomes +8 MW, and the total injected power of MG5 becomes +15 MW. Based on these new boundary conditions, the power flow calculation engine recalculates the predicted power flow distribution of all lines in the entire network, obtaining the predicted transmission power values. Subsequently, for each transmission line connected to or significantly affected by these microgrids, such as line L28 with a predicted power flow of 185 MW, the system extracts the maximum allowable transmission power of line L28 from the grid asset database, for example, 200 MW. This maximum value is based on the line's static... The thermal stability limit is set, and then the percentage of the predicted transmission power value to the maximum allowable value is calculated, i.e., 185 divided by 200, which gives 92.5%. Since this percentage does not exceed 100%, the transmission capacity of line L28 is considered safe. However, for another line L36, if its predicted power flow is 160 MW, while its maximum allowable value is only 150 MW, then its load factor ratio is 106.7%, which exceeds 100%. The system identifies that the line has an over-limit risk. At this time, the system will start a power redistribution algorithm. This algorithm aims to minimize the adjustment amount and finely adjusts the power allocation value of the microgrid unit (such as MG5) that caused the line to exceed the limit. For example, the injected power of MG5 is reduced from 15 MW to 13.5 MW. Then the power flow calculation is run again until the transmission power of all lines is within the allowable range, completing the frequency support power adjustment and allocation under the constraints, and generating a microgrid dynamic frequency support scheduling scheme.

[0175] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. An optimal decentralized coordination control method for microgrids in new energy power systems, characterized in that, The method comprises the following steps: Based on the online acquired power grid topology and real-time line flow data, the key transmission lines are virtually disconnected one by one for verification, and a whole-network cascading risk quantification index is established; Based on the whole-network cascading risk quantification index, a numerical comparison is made with a safety threshold, when the whole-network cascading risk quantification index exceeds the threshold, the initial fault line with the highest contribution to the whole-network cascading risk quantification index is identified, a vulnerable line risk contribution list is generated, and the power injection or absorption adjustment amount executed by the micro-grid group is calculated and determined according to the vulnerable line risk contribution list, and a micro-grid group preventive power flow diversion instruction set is acquired; In the operating state of executing the micro-grid group preventive power flow diversion instruction set, each micro-grid unit continuously monitors to generate a cluster-in overrun state information set, and the cluster-in overrun state information set is compared with the instruction update amount issued by the cluster coordinator and the preset threshold to generate a hierarchical event trigger activation signal; Based on the hierarchical event trigger activation signal, each cluster coordinator starts collaborative optimization calculation to generate a cluster-level optimal control sequence, the cluster-level optimal control sequence is checked and distributed to establish a micro-grid dynamic frequency support scheduling scheme; The acquisition step of the micro-grid dynamic frequency support scheduling scheme is: Based on the cluster-level optimal control sequence, the planned calling power and the planned calling time period of the energy storage device of each micro-grid unit in the cluster-level optimal control sequence are extracted, and the predicted state of charge after the implementation of the planned calling power is compared with the upper and lower limit values of the allowed state of charge interval based on the current actual state of charge value and the allowed state of charge interval of the corresponding micro-grid unit, the micro-grid units whose state of charge satisfies the interval constraint condition are screened, and a control sequence after the state of charge constraint check of the energy storage is generated; Based on the control sequence after the state of charge constraint check of the energy storage, the planned power adjustment amplitude of the generator set in each micro-grid unit in the sequence is acquired, the actual ramping rate of the generator set is calculated based on the allowed range value of the ramping rate of the corresponding generator set and the planned adjustment time interval value, and the actual ramping rate is compared with the allowed range, the adjustment items of the micro-grid units that do not exceed the ramping rate constraint condition are retained, and a control sequence after the ramping rate constraint check of the generator is generated; Based on the control sequence after the ramping rate constraint check of the generator, the predicted line transmission power value after the frequency support action of each micro-grid unit in the sequence is retrieved, the maximum allowed transmission power value of the corresponding line is extracted, and the percentage of the transmission power value to the maximum allowed value is calculated, if the percentage exceeds 100%, the power distribution value of the corresponding micro-grid unit is adjusted to meet the line transmission capacity constraint condition, the frequency support power adjustment distribution under the constraint condition is completed, and a micro-grid dynamic frequency support scheduling scheme is generated.

2. The optimal decentralized coordination control method for microgrid in new energy power system according to claim 1, characterized in that, The acquisition step of the whole-network cascading risk quantification index is: Based on the online collected power grid topology information and the current state of each transmission line, the power flow value, any line in the key transmission line set is selected for disconnection operation simulation, the power flow change value of the remaining lines is recorded, and the total power flow amplitude under the accident scene is calculated combined with the power flow value before disconnection, to generate the disconnection line number and the post-accident line power flow table; According to the disconnection line number and the post-accident line power flow table, the power flow increment and the initial load power before and after the accident are extracted, and the load rate of each line under the current disconnection scene is calculated, and the proportion of the initial power flow in the total power flow of the whole network is extracted, to generate the risk contribution value set of all lines under the disconnection scene; According to the risk contribution value set of all lines under the disconnection scene, the whole network cascading risk quantitative index is calculated.

3. The optimal decentralized coordination control method for microgrid in new energy power system according to claim 1, characterized in that, The acquisition step of the fragile line risk contribution list is: Based on the whole network cascading risk quantitative index, the pre-acquired safety threshold value is set, the whole network cascading risk quantitative index is compared with the safety threshold value, whether the whole network cascading risk quantitative index exceeds the safety threshold value is judged, and the risk overrun judgment result is generated; According to the risk overrun judgment result, if the whole network cascading risk quantitative index exceeds the safety threshold value, the disconnection line number and the post-accident line power flow table are traced back, each disconnection line scene is traversed, and the corresponding risk contribution value set is extracted, the single line risk contribution value in each risk contribution value set is accumulated and summed, and the risk contribution cumulative value of each line fault scene is formed; Based on the risk contribution cumulative value of each line fault scene, the risk contribution cumulative value is sequentially sorted, the line with the highest risk contribution cumulative value is selected as the initial fault line, and the initial fault line and the corresponding risk contribution cumulative value are recorded, and the fragile line risk contribution list is formed.

4. The optimal decentralized coordination control method of microgrid in new energy power system according to claim 1, characterized in that, The acquisition step of the micro-grid group preventive power flow dredging instruction set is: According to the fragile line risk contribution list, the risk contribution cumulative value and the line number corresponding to each line in the fragile line risk contribution list are extracted, the line number is sorted according to the risk contribution cumulative value, and the power injection or power absorption adjustment value required for the corresponding micro-grid is determined according to the micro-grid distribution of each line, to generate the micro-grid group power adjustment value set; Based on the micro-grid group power adjustment value set, the power injection or power absorption adjustment value corresponding to each micro-grid unit is mapped to the executable power flow dredging instruction one by one, and the power flow dredging instruction is distributed to the corresponding micro-grid unit with the micro-grid unit number as the index, to form the micro-grid group preventive power flow dredging instruction set.

5. The optimal decentralized coordination control method for microgrid in new energy power system according to claim 1, characterized in that, The acquisition step of the intra-cluster overrun state information set is: Based on the execution state of the micro-grid group preventive power flow dredging instruction set, the local monitoring device of each micro-grid unit continuously collects the frequency deviation value of the current micro-grid unit, the state of charge value of the energy storage device, and the operation state change value of the neighbor micro-grid unit, and compares them with the corresponding preset safety operation threshold value one by one, records the value entries and micro-grid unit numbers that exceed the preset safety operation threshold value, and forms the intra-cluster overrun state information set.

6. The optimal decentralized coordination control method of microgrid in new energy power system according to claim 1, characterized in that, The acquisition step of the hierarchical event trigger activation signal is: Based on the set of out-of-limit state information in the cluster and the instruction update amount issued by the cluster coordinator, the frequency deviation value, the energy storage state of charge change value and the neighbor cluster power exchange change value of each micro-grid unit are read one by one, and the frequency adjustment target value, the energy storage charge and discharge plan value and the power exchange expectation value of the micro-grid unit in the same time period are matched and extracted, the absolute value of the difference between the state deviation and the instruction target value is calculated respectively, and then compared with the corresponding frequency deviation threshold, state of charge threshold and power exchange deviation threshold, the micro-grid unit identification and deviation type of any absolute value exceeding the set threshold are screened, and the hierarchical event trigger activation signal is generated.

7. The optimal decentralized coordination control method of microgrid in new energy power system according to claim 1, characterized in that, The acquisition step of the cluster-level optimal control sequence is: According to the hierarchical event trigger activation signal, the individual dynamic adjustment urgency index of each micro-grid is calculated; Based on the individual dynamic adjustment urgency index of each micro-grid unit, the state characteristics of the micro-grid unit with the index value in the top 30% interval are retrieved in order, and the adjustable power interval, response time delay and current energy storage state of charge are extracted, combined with the load change amount and the expected adjustment effect of each type of execution action, the optimal control action set is screened and assembled into a continuous adjustment sequence, and the cluster-level optimal control sequence is generated.

Images